Team Kairos: Aiding Astronauts

The capstone project at Carnegie Mellon's HCII pairs teams of students in the master's program with an external client who challenges them to research, design, and prototype a user-centered system over eight months. The spring semester is spent in the research phase, during which students are also enrolled full-time in other electives in Pittsburgh. The summer semester is wholly devoted to design and development.

My client, the NASA Ames Research Center, brought to my team a missive to reimagine the way schedules procedures are presented to astronauts as they execute tasks in space. As Research Lead for the full duration of the eight-month project, I led the efforts to narrow our research scope and conduct contextual inquiries and interviews in the spring, and then led usability testing on our prototype designs in the summer. Full details are below.

Team:
Samia Ahmed (Design Lead), Diana Chen (Project Manager), Cara Esten Hurtle (Technical Lead), and Joe Medwid (UX Lead), all fellow master's students in human–computer interaction at Carnegie Mellon University.

Duration:
Eight months (January 2012–August 2012)

Skills used:
Literature review, competitive analysis, field research, contextual inquiry, affinity diagramming, sketching, speed dating, Balsamiq, OmniGraffle, iOS programming, usability testing, expert reviews, think aloud protocol

Full research report:
Download!

Website:
Team Kairos

Background research

After conducting a literature review to learn more about our domain and about general challenges with complex procedure execution, we set off for Johnson Space Center in Houston, TX, where we observed first-hand how mission operations are conducted and spoke with many employees about the challenges astronauts currently face.

Although user-centered research naturally involves speaking with users, and ideally observing them in situ, there wasn't room in the project budget to send us to space. In lieu of observing astronauts executing procedures, we compiled a list of research foci we wished to observed in other analogous domains. Ultimately, we spoke to professionals in stage management, automotive repair, surgery, and construction.

Team Kairos in front of the Deep Space Habitat.
Esten learns about cockpit prototypes for the Multi-Purpose Crew Vehicle.
An affinity diagram we constructed.

Research findings

After hours of interviews and observation and thousands of post-it notes, we found our research described eight high-level findings which can be further divided into considerations—facts about mission operations over which we have little control—and insights—ideas which we consider to be actionable. These have guided our design process throughout the summer and are listed below.

Considerations

In high-pressure operations, staff must provide assertive direction and moral support

Being an astronaut is highly stressful, and it's a position that you only obtain after years of training. We found that professional respect, however, can make it difficult for ground staff to provide appropriate criticism or to intervene when a crewmember needs help. Similarly, we found that the bond between ground staff and astronaut is especially strong, and this affords a great opportunity to reduce the stress of high-pressure operations through moral support.

Future missions must account for intermittent ground–crewmember communication

In a discovery that shouldn't surprise anyone, NASA's desire to explore further into space threatens the tight coupling of their ground–crewmember support strategy. Communication delays of a few seconds to the moon were never an issue, but a mission to Mars will involve constant delay. We found that in simulated missions NASA has run with communications delay, text becomes the most viable medium for remote communication, and that users will frequently switch tasks while waiting for a response.

Inventory management issues can delay procedure execution

Physical tool-finding is quite hard on ISS at the moment, with more than 30,000 distinct items aboard. The current process crewmembers use to look up the locations for these tools before beginning a procedure is cumbersome. Additionally, because of variable compliance by crew in updating these data, sometimes locations are inaccurate. With a highly-packed schedule, even a slight delay can cause a waterfall of changes.

Individual crewmembers read and understand procedures at various levels of granularity

To anyone besides the procedure authors and executors, most space procedures are semantically meaningless—an alphabet soup of acronyms, complex mechanical devices, and regulations that probably apply to a device of which there may only be one other copy in existence besides the one in space. Additionally, astronauts are generalists, not specialists, and one user may not be as familiar as the next with a given piece of equipment. These procedure quickly get detailed as a result.

Insights

Critical contextual information is obscured by items of less immediate concern

Currently, users are given a long-scale schedule of their entire day, with a horizontal swimlane design. This design can make it very difficult to see anything independent of time, such as any information about a scheduled activity if it's very short. Additionally, procedures themselves do very little in the way of hierarchy to communication changes from the last time a crewmember performed that procedure.

Methods that encourage memory recall can support consistent procedure execution

Crewmembers perform a wide variety of tasks very infrequently, and it can be tricky for them to remember their training. In our analogous domains, we found that notetaking and explicit checklists can be used to counter memory loss to great effect—but these mediums aren't integrated into NASA's current procedure system.

Existing procedure support systems do not prioritize users' most pressing needs

Ground staff likes to know exactly what their crewmembers are doing at all times, but the current software system provides only a cumbersome way for crewmembers to communicate their progress in or within a task. There are great opportunities for passive data collection in a future redesign.

Instructive systems should not make superfluous demands on users' cognitive load

Cognitive load, already at a premium with complex procedures to be executed, shouldn't be further taxed by a confusing software system that makes even switching between users a multi-step process. Greater visual hierarchy within procedures and more natural interaction paradigms should aid in minimizing additional cognitive load from an instructive system.

What's Next

I am currently working on-site at NASA Ames Research Center to complete this project in early August. By then we will have developed the visions and findings communicated in our report into a full, working iPad prototype. Because of my experience in iOS development, I have been doing a great deal of development in addition to my role as research lead in the summer.